Deep Learning for Heart Rate Estimation From Reflectance Photoplethysmography With Acceleration Power Spectrum and Acceleration Intensity

Hoon Ko,Heewon Chung,Jinseok Lee,Hooseok Lee

doi:10.1109/access.2020.2981956

Abstract

Objective: A wearable reflectance-type photoplethysmography (PPG) sensor can be incorporated in a watch or band to provide instantaneous heart rates (HRs) with minimum inconvenience to users. However, the sensor is sensitive to motion artifacts (MAs), which results in inaccurate HR estimation. To address this problem, we propose a new neural network for deep learning to ensure accurate HR estimation even during intensive exercise. Methods: We propose a new deep neural network based on multiclass and non-uniform multilabel classification for HR estimation. It comprises of two convolutional layers, two long short-term memory (LSTM) layers, one concatenation layer, and three fully connected layers including a softmax. The proposed model feeds the power spectra from the PPG and acceleration signals along with the acceleration intensity to the input layer. We also present a new scheme to evaluate the loss value by modifying the true HR value into a Gaussian distribution. Results: We used 48 training datasets and evaluated 23 isolated testing datasets. The proposed model exhibited average absolute error of less than 1.5 bpm for all the training and test datasets—1.09 bpm for the training dataset and 1.46 bpm for the test dataset. Conclusion: The proposed model outperforms the state-of-the-art methods for accurate estimation of HR. Significance: It precisely estimates the HRs with robustness even during intensive physical exercise, as evidenced by the accuracy when PPG signals are severely corrupted by MAs.

Highlights

Reflectance-type photoplethysmography (PPG) sensor measures intensity changes in the light reflected from skin, providing PPG signals that represent the changes in the arterial blood volume between the systolic and diastolic phases of a cardiac cycle
FROM TRAINING AND TEST SETS Based on our proposed model, we found that the resultant average absolute errors (AAEs) and ARE values were 1.09 bpm and 0.92% for the training dataset (n = 48), and 1.46 bpm and 1.23% for the test dataset (n = 23), respectively
The proposed model was sequentially structured with a 2D convolutional layer, a 1D convolutional layer, and a fully connected layer, which were incorporated for motion artifacts (MAs) cancelation

Summary

Introduction

Reflectance-type photoplethysmography (PPG) sensor measures intensity changes in the light reflected from skin, providing PPG signals that represent the changes in the arterial blood volume between the systolic and diastolic phases of a cardiac cycle. The sensor is sensitive to motion artifacts (MAs), which originate from pressure and movement applied on the wrist on which the PPG sensor is worn. The MAs eventually result in inaccurate HR estimation. Years ago, Zhang et al shared the datasets containing simultaneously measured acceleration and PPG signals during exercise [1], which has prompted research on MA cancelation in PPG sensors using acceleration signals. Most state-of-the-art methods estimated the HR using the two main stages of MA cancelation and HR tracking. For MA cancelation, they considered the power spectrum from the simultaneously measured acceleration signal as motion artifacts (MAs), and removed or attenuated the power from the PPG power spectrum.

Methods

Results

Conclusion